Add C and CXX API for homophone replacer (#2156)

2025-04-27 22:09:13 +08:00
parent f64c58342b
commit e51c37eb2f
13 changed files with 586 additions and 1 deletions
--- a/c-api-examples/sense-voice-with-hr-c-api.c
+++ b/c-api-examples/sense-voice-with-hr-c-api.c
@@ -0,0 +1,99 @@
+// c-api-examples/sense-voice-with-hr-c-api.c
+//
+// Copyright (c)  2024-2025  Xiaomi Corporation
+
+//
+// This file demonstrates how to use SenseVoice with sherpa-onnx's C API
+// with homophone replacer.
+// clang-format off
+//
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+// tar xvf sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+// rm sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+//
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/hr-files/dict.tar.bz2
+// tar xf dict.tar.bz2
+//
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/hr-files/replace.fst
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/hr-files/test-hr.wav
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/hr-files/lexicon.txt
+//
+// clang-format on
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sherpa-onnx/c-api/c-api.h"
+
+int32_t main() {
+  const char *wav_filename = "./test-hr.wav";
+  const char *model_filename =
+      "./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/model.int8.onnx";
+  const char *tokens_filename =
+      "./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/tokens.txt";
+  const char *language = "auto";
+  const char *provider = "cpu";
+  int32_t use_inverse_text_normalization = 1;
+
+  const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
+  if (wave == NULL) {
+    fprintf(stderr, "Failed to read %s\n", wav_filename);
+    return -1;
+  }
+
+  SherpaOnnxOfflineSenseVoiceModelConfig sense_voice_config;
+  memset(&sense_voice_config, 0, sizeof(sense_voice_config));
+  sense_voice_config.model = model_filename;
+  sense_voice_config.language = language;
+  sense_voice_config.use_itn = use_inverse_text_normalization;
+
+  // Offline model config
+  SherpaOnnxOfflineModelConfig offline_model_config;
+  memset(&offline_model_config, 0, sizeof(offline_model_config));
+  offline_model_config.debug = 1;
+  offline_model_config.num_threads = 1;
+  offline_model_config.provider = provider;
+  offline_model_config.tokens = tokens_filename;
+  offline_model_config.sense_voice = sense_voice_config;
+
+  // Recognizer config
+  SherpaOnnxOfflineRecognizerConfig recognizer_config;
+  memset(&recognizer_config, 0, sizeof(recognizer_config));
+  recognizer_config.decoding_method = "greedy_search";
+  recognizer_config.model_config = offline_model_config;
+  recognizer_config.hr.dict_dir = "./dict";
+  recognizer_config.hr.lexicon = "./lexicon.txt";
+
+  // Please see
+  // https://colab.research.google.com/drive/1jEaS3s8FbRJIcVQJv2EQx19EM_mnuARi?usp=sharing
+  // for how to generate your own replace.fst
+  recognizer_config.hr.rule_fsts = "./replace.fst";
+
+  const SherpaOnnxOfflineRecognizer *recognizer =
+      SherpaOnnxCreateOfflineRecognizer(&recognizer_config);
+
+  if (recognizer == NULL) {
+    fprintf(stderr, "Please check your config!\n");
+    SherpaOnnxFreeWave(wave);
+    return -1;
+  }
+
+  const SherpaOnnxOfflineStream *stream =
+      SherpaOnnxCreateOfflineStream(recognizer);
+
+  SherpaOnnxAcceptWaveformOffline(stream, wave->sample_rate, wave->samples,
+                                  wave->num_samples);
+  SherpaOnnxDecodeOfflineStream(recognizer, stream);
+  const SherpaOnnxOfflineRecognizerResult *result =
+      SherpaOnnxGetOfflineStreamResult(stream);
+
+  fprintf(stderr, "Decoded text: %s\n", result->text);
+
+  SherpaOnnxDestroyOfflineRecognizerResult(result);
+  SherpaOnnxDestroyOfflineStream(stream);
+  SherpaOnnxDestroyOfflineRecognizer(recognizer);
+  SherpaOnnxFreeWave(wave);
+
+  return 0;
+}